Having a mortgage.

SuperTechBio

At Microsoft Build yesterday, Satya Nadella released Microsoft Discovery, the framework for building AI scientists that can think and reason. What was the core? It wasn't an LLM. It's the knowledge graph and the ontology. Sitting underneath every AI agent that can reason,

"Through AXL, Toronto can be the Silicon Valley of the 21st century." 🚀 LFG

One of the greatest barriers to human progress is knowledge transfer. The faster we can transmit bits of information to one another, the sooner we can build the future.

I used text-to-voice on all the rejection emails and made an mp3. I listen to it while I work out. Infinite energy. My PR growth now matches my ARR growth.

🔥Just wrapped up an incredible three days in Montreal at the AACR (AACR ) Special Conference on Artificial Intelligence and Machine Learning — the first AACR event fully dedicated to exploring how AI is reshaping cancer research and care. I was deeply honored to serve as one of

🎤 I had the pleasure of chatting with Christopher Li , CEO & Co-Founder @ BioBox Analytics Bio Box provides a knowledge and reasoning infrastructure for biopharma R&D, helping scientists scale up hypothesis generation and testing in early drug discovery. Chris is a top founder

Novel target biology will be the battleground for the next decade.

Pitching angel groups is a weird experience. Maybe sector specific. But when founders get unsolicited "business advice" from someone who has never operated or built a company or ppl start debating the founder around competition when you're not even in the industry, don't be

Building a globally competitive US biotech in 2025: > Focus on novel biology why? —> The Chinese biotech VC ecosystem struggles to fund novel biology (actively disincentivizes going after non clinically validated targets) —> The US biotech ecosystem struggles to execute on

Switching from building the software to building the business as a founder is a very hard transition, especially if you are technical. So much satisfaction when you get to lock in and solve a problem, ship it, and see results. You have full control. It's like a drug. Then, at

Saying your "AI agent" is trained on every paper in scientific literature, pre-print, and abstract is fundamentally equivalent to feeding it garbage.

In drug discovery, your data is your moat. Yet too many “graph” vendors sell you access to their knowledge graph and then require you to hand over your proprietary data just to make it usable. The catch? ❌ You don’t actually own the graph. ❌ You don’t control how your data is

Saw a Reddit thread from a comp bio PhD student asking if 4-6 hours of deep work is enough because their roommate is putting in 18-20 hour days. Then an onslaught of comments shaming the “grind”. Yea… guaranteed ngmi for all them.

Someone please explain how this makes sense….

Biology is a graph. Any other representation is a lossy approximation that handicaps your ability to do science at scale. I recently wrote a blog post outlining my arguments for this here. blog.biobox.io/p/biology-is-a…

This is an amazing leap forward. Congrats Sam Rodriques !

😂 I can't wait to hear how drug makers feel about (6). Zero chance OAI will get that deal approved.